Documentation

Welcome to SigOpt’s developer documentation. If you have a question you can’t answer, feel free to contact us!

Prior Object

A prior belief distribution of a Parameter. By default, SigOpt assumes the parameter to be uniformly distributed. SigOpt allows users to specify the following distributions on double parameters. Refer to the Prior Beliefs documentation for more information.

Beta Prior

SigOpt implements the generalized Beta distribution, which normalizes the Beta distribution to the parameter bounds.

Normal Prior

SigOpt implements the truncated Normal distribution, which truncates the Normal distribution to the parameter bounds.

Fields

KeyTypeValue
meanfloatOnly supported for normal prior. Mean of the truncated Normal distribution.
namestringThe name of the prior distribution. Currently only supports beta and normal priors.
scalefloatOnly supported for normal prior. Standard deviation of the truncated Normal distribution.
shape_afloatOnly supported for beta prior. Shape parameter α of the Beta distribution.
shape_bfloatOnly supported for beta prior. Shape parameter β of the Beta distribution.

Examples

Beta Prior

{
  "name": "beta",
  "object": "beta_prior",
  "shape_a": 2,
  "shape_b": 4.5
}

Normal Prior

{
  "mean": 0.5,
  "name": "normal",
  "object": "normal_prior",
  "scale": 1.2
}